CT measures of femoral and tibial version and rotational position of femoral and tibial components of knee replacements: limitations in reliability and suitability for routine clinical practice

Objectives Rotational malalignment of knee replacements as measured on CT is understood to be associated with poor outcomes. The aim of this study is to measure the inter-rater and intra-rater reliability of measures of femoral and tibial version in the native arthritic knee and postoperative TKR component position using CT. Methods Eighty patients underwent CT of the knee before and after total knee replacement. Preoperative femoral and tibial version and component rotation were independently measured by two musculoskeletal radiologists. Results Mean differences between and within raters were small (< 1.6°). Maximum 95% limits of agreement for inter-rater and intra-rater comparisons were 8.1° and 7.6° for preoperative femoral version, 9.0° and 7.9° for postoperative femoral rotation, 26.0° and 20.5° for preoperative tibial version, and 24.9° and 23.6° for postoperative tibial rotation respectively. Postoperative ICCs varied from 0.68 to 0.81 (lower 95% CI:0.55–0.72) for both intra- and inter-rater comparisons. Preoperative ICCs were lower: 0.55–0.75 (lower 95% CI:0.40–0.65). Conclusion The lower 95% confidence level for ICC of version and rotational measurements using the Berger protocol of TKRs on CT are all less than 0.73 and that the normal range of differences between observers is up to 9° for the femoral component and 26° for the tibial component. This suggests that CT measurements derived from the Berger protocol may not be consistent enough for clinical practice. Key Points • CT is commonly used to measure the rotational profile of knee replacements in symptomatic patients using the Berger protocol. • The limits of agreement for both femoral and tibial component rotation are wide even for experienced observers. • CT measurements of the rotation of knee arthroplasty are not reliable enough for routine clinical use.